Research

Research themes

My research projects lie in three areas under language acquisition:

  1. Child multilingual acquisition involving Mandarin, Cantonese and English: Simultaneous and sequential development of two or more languages involving Chinese and English in infants and preschool children, with monolingual children and their input as baselines
  2. Heritage Chinese in overseas communities: Early development, ultimate attainment and attrition of the grammatical system in heritage language speakers of Chinese across the life span, with a special interest in the transition period of heritage bilingualism when the societal majority language gradually replaces the heritage languages as the dominant language of the bilingual speaker
  3. Adult language acquisition (L2 Chinese, L3 French/German): Functional categories in adult second language (L2) speakers of Chinese such as aspect markers, focus-sensitive particles and grammatical constructions subject to multi-level interface conditions (e.g. ba, bei, topic and cleft); a recent addition to this line of research is a Co-PI role in a collaborative project taking a predictive personalized approach to enhancing L3 learning of French and German (PC: Gangyi Feng)

I explore these topics through a combination of corpus, experimental and statistical methods.

Research grants

Research Grants Council (RGC), HKSAR

  • ​As PI: “Input dilution and supplementation in monolingual and bilingual siblings” (General Research Fund/GRF, Research Grants Council of Hong Kong/RGC, Project No. 14623425, 2026-2027, HK$973,000), RGC intro page
  • ​As Co-PI: “A predictive personalization approach to enhance foreign language learning and teaching” (Young Collaborative Research Grant/YCRG, Project No.: C4001-23Y, 2024-2027), PC: Gangyi Feng
  • ​As PI: “Input and caretaker proficiency in early bilingual development: mothers, helpers and toddlers” (Early Career Scheme/ECS, Research Grants Council of Hong Kong/RGC, Project No. 21604522, 2023-2024, HK$530,000), RGC intro page
  • ​As PI: “Input and experience in early trilingual development” (General Research Fund/GRF, Research Grants Council of Hong Kong/RGC, Project No. 14615820, 2021-2023, HK$977,820), RGC intro page

CUHK-internal (as PI)

  • ​“Optimizing caretaker input for Chinese-learning children in multilingual acquisition contexts”, Strategic Impact Enhancement Fund, University Research Committee, 2025-2027
  • “Learning from ‘digital babies’: from predictive models to language input planning”, Strategic Seed Funding for Collaborative Research Scheme, University Research Committee, 2025-2027
  • “Cantonese-English bilingual children’s reference production: An eye-tracking study”, Direct Grant for Research, Faculty of Arts, 2024-2025, with Jiangling Zhou
  • “Quantity and quality of input in early bilingual development”, Direct Grant for Research, Faculty of Arts, 2019-2020
  • “Trilingualism in early childhood: production and comprehension”, Direct Grant for Research, Faculty of Arts, 2016-2017

Research staff and students (updated Aug 2025)

Postdoctoral fellows

  • Dr SHANG, Mengyao (Melody) 尚孟堯 (2024-): Melody earned her MA & PhD in Linguistics from CUHK. Her research employs corpus-based and experimental methods to investigate the acquisition of complex structures in monolingual and bilingual children. In her postdoctoral work, she is also examining the critical role of language input in child language development. Contact: mengyaoshang@cuhk.edu.hk
  • Dr LIU, Zhenting (Sally) 劉真廷 (2025-): Zhenting received her MA and Ph.D. in Linguistics from CUHK. Her research interests focus on speech perception in infants and the phonological development of children growing up in multilingual contexts. Contact: ztliu@cuhk.edu.hk

PhD students

  • Miss LIANG, Yuqing (Lilian) 梁雨晴 (as Supervisor in 2021-2023; as External Co-Supervisor since 2023, CityU): Yuqing is a final-year Ph.D. student. Her research interest focuses on input and outcomes in early bilingual and trilingual development. Contact: yuqiliang7-c@my.cityu.edu.hk
  • Miss LIU, Jingyao (Jane) 劉婧瑤 (as Supervisor in 2021-2023; as External Co-Supervisor since 2023, CityU): Jingyao is a final-year Ph.D. student. Her dissertation, currently awaiting oral defense, examines the naturalistic production of only-type exclusive focus in both child-produced and child-directed speech. Her research interests include language acquisition in monolingual and bilingual children, and the role of input in language development. Contact: jingyliu27-c@my.cityu.edu.hk
  • Miss ZHANG, Xuening (Tracy) 張雪寧 (as Supervisor in 2022-2023; as External Co-Supervisor since 2023, CityU): Xuening is a fourth-year Ph.D. student. She received her B.A. in English Language and Literature from Shandong University. Her research interests are child language development and multilingual acquisition. Contact: xuenzhang2-c@my.cityu.edu.hk
  • Miss CHEN, Yuxuan (Chloe) 陳雨璇 (2024-, CUHK): Yuxuan received her M.A. in Linguistics from the University of Hong Kong. Her research focuses on language acquisition and bilingualism, particularly in bilingual children’s acquisition. Contact: chenyuxuanchloe@link.cuhk.edu.hk
  • Miss NIE, Jiaqi (Nicole) 聶嘉淇 (2024-, CUHK): Nicole earned her B.A. in English from Guangdong University of Foreign Studies and her M.A. in Linguistics from the University of Hong Kong. Her research focuses on maternal input, both native and non-native, and its relationship to bilingual language development in young children. Contact: jiaqi.nie@link.cuhk.edu.hk
  • Miss ZHANG, Fan (Ashley) 張帆 (2025-, CUHK): Fan obtained her B.A. in English from Northeastern University and her M.A. in Linguistics from Lanzhou University. Her principal interests are in the fields of cross-linguistic influence and second/third language acquisition via experimental approaches. Contact: ashleyzfan@link.cuhk.edu.hk

Research assistants

  • Miss CAO, Yue (Jenn) 曹玥
  • Miss LIANG, Mingwei (Ellen) 梁明蔚
  • Miss YE, Peng (Agnes) 葉芃: Agnes obtained her M.A. in Linguistics from the Chinese University of Hong Kong and her B.Econ in Finance from Jilin University. Her research interests lie in multilingualism and the early acquisition of verbs and tense/aspect marking.
  • Miss CHAN, Ashley (part-time)

Former lab members

Behind-the-scenes stories (updated Aug 2025)

  • Focus constructions and their acquisition by adults and children

This series of research originated from my PhD project supervised by Professor Boping Yuan at the University of Cambridge, funded by the former Overseas Research Scholarship (ORS) and the Cambridge Overseas Trust (2008-2012). Initially we wanted to examine the formal features attached to the shi..de focus construction (e.g., focus-driven A’-movement), but we ended up with a larger set of features beyond narrow syntax including temporal semantics and discourse presuppositions, which, we believe, provided interesting evidence to address both the Interface Hypothesis and the Feature Reassembly Hypothesis. We found that across the features, the discourse feature presents the greatest challenge to English-dominant learners, but the nature of the challenge is different between the adult L2 learners and heritage speakers. Our analysis of the cleft construction was inspired and influenced by the work of Waltraud Paul and Lisa Cheng in 2008, although we did not test the most interesting “shi…de cleft proper” in the V-de-O form (e.g., 小李是昨天吃的蛋糕), because in our pilot studies, lower-proficiency learners did not accept this sentence as a cleft in general. (Instead, they interpreted this as a relative clause as in Lee was the cake that was eaten yesterday, ha!). In this study, we postulated, rather than directly examined the role of prosodic stress in the focus cleft, which is obviously non-trivial. The shi…de cleft still presents intriguing puzzles for both theoretical analyses and acquisition studies to this very day.

Outputs: Some observations of the temporal semantic properties of the shi…de cleft were reported in Chinese in Essays on Linguistics (《语言学论丛》) in 2011. After sharing some of the main findings in the GASLA12 proceeding (Mai, 2013), in two subsequent publications, we reported the findings of the adult L2 learners within the framework of Feature Reassembly (Mai & Yuan, 2016), and the findings of the adult heritage learners enriched with child L1 data (Mai & Deng, 2019). The findings of this project have been cited in textbooks such as Second Language Learning Theories (Mitchell et al., 2019), How Second Languages Are Learned: An Introduction (Hawkins, 2018), and The Acquisition of Heritage Languages (Montrul, 2016). Another milestone in this rewarding journey with adult L2 Chinese learners is my review article on grammatical development and processing in L2 Chinese in Second Language Research (Mai, 2016).

Current follow-ups: My PhD student Jingyao Liu is working on the acquisition of two focus particles (only in English, zhi(you) in Mandarin) in monolingual and bilingual children and we’re finally factoring in the prosodic features in addition to the grammatical ones in our analysis!

  • The ba-construction and heritage Chinese across the life span

Experiments are costly and fillers are necessary. In my PhD project on the shi…de cleft, I had included, as fillers, carefully-designed items testing several syntactic, semantic and discourse properties of the world-famous ba-construction. I had little idea that I had planted seeds which would grow into an important line of my research in the next few years. In the ba-fillers, we found that the less commonly found “causative ba-construction” depicting causative events with an inanimate Cause and an animate Affectee (e.g., 那杯酒把小李喝醉了) was categorically rejected by the heritage speakers (who were L1 speakers of Mandarin) even when lexical scaffolding and appropriate contexts had been provided, whereas the L1 Chinese controls from Beijing readily accepted them. We reported this in a conference paper in GALA12 (Mai & Yip, 2015). At the same time, we looked at the zoeng-construction (a structural equivalent of ba) in heritage speakers of Cantonese in English-dominant communities in an elicited production experiment and found that they produced the zoeng-construction with low production rates yet high accuracy rates (Mai, Kwan & Yip, 2018). We were intrigued and wanted to find out what had happened at earlier stages of heritage language development so we began to look at younger heritage speakers.

School-age speakers: We began by looking at school-age heritage children in the UK. We collaborated with Lucy Zhao and collected bilingual frog story narratives from both heritage children and their Chinese-speaking parents in England and compared the ba-constructions produced by the children and the parents, with existing frog stories told by children and adults in Beijing in CHILDES as references. We found that the children were very conservative and used a smaller set of lexical nouns and verbs in the ba-sentences than their parents, and their choice of those lexical items were apparently influenced by their equivalents in English. Interestingly, whether the parents used a wide range of words to tell the stories (lexical diversity) was positively correlated with their children’s production rate of ba-utterances in their frog stories (Mai, Zhao & Yip, 2022). Using the same dataset, we published two more studies examining two complex verbal structures in Mandarin: the resultative verb compound (RVC, Shang, Zhao, Yip & Mai*, 2024) and the pre/post-verbal prepositional phrases headed by zai (zai-PPs, Deng & Mai*, 2024). All three studies suggest that for high-proficiency heritage speakers of isolating languages like those in our studies, the lexicon (rather than core syntax) is likely to be the source of their protracted development.

Preschoolers: At the same time when we conducted the cross-sectional study in Sheffield, we launched another project documenting the bi/trilingual development of American-born Chinese children at the pre-school age longitudinally, namely the “Child Heritage Chinese Corpus/CHCC”. Those children were exposed to their heritage language (Chinese = Mandarin and/or Cantonese) at home, and the societal majority language (English) mostly at school (“one context-one language”). We have deposited naturalistic adult-child play data from three of the children to CHILDES: Luna, Avia and Winston. To download corpus (transcripts and audio recordings), visit CHILDES info page: CHCC. To know more about the corpus, visit CBRC info page: CHCC. More data are being collected and processed, and several manuscripts are forthcoming.

  • Early multilingual acquisition in Chinese-speaking households and input-outcome relations

I relocated to Hong Kong after my PhD and joined the Childhood Bilingualism Research Centre at CUHK as a postdoctoral research fellow working closely with Virginia Yip, who introduced me to this amazing field of childhood bilingualism and demonstrated to me the great potential of testing central hypotheses in language acquisition with multilingual children in Hong Kong. When my son Leo arrived, my family were determined to raise him trilingually in Cantonese, Mandarin and English using existing language resources of the caretakers at home. It didn’t take long for us to hammer on the optimized trilingual input package for Leo: we incorporated “one day-one language” and “one parent-one language” input models and adhered to it for three years. Leo successfully developed productive trilingualism and appropriate code-switching abilities by 2;11. This longitudinal study was documented in the Leo Corpus in CHILDES. Leo’s development in Mandarin and code-mixing patterns are reported in Mai, Wu and Yip (2021), and his overall development in the three languages (especially Cantonese and English) and input-outcome relations are discussed in Mai and Yip (2022). More about Leo and his language development can be found on the CBRC info page: Leo.

One of the many perks of being a member of Childhood Bilingualism Research Centre (CBRC) is the opportunity to be introduced to passionate and like-minded researchers at similar stages of personal and career development. I met Xiangjun Deng and Jiangling Zhou here. I have had the honour to witness the birth of the widely-cited Tong Corpus, and collaborated with Xiangjun and Jiangling on projects on ba and bei in Mandarin monolingual children (Deng, Mai & Yip, 2018) and referential expressions by bilingual children in Hong Kong (Zhou, Mai & Yip, 2021) and Singapore (Zhou, Mai, Cai, Liang, Yip, 2022). Look out for more joint papers by us!

Current follow-ups: Thanks to Leo and the numerous students and colleagues who worked on the Leo project, we presented our case study and obtained funding from the Research Grants Council of HKSAR to investigate input effects in bilingual and trilingual children in larger sample sizes. We also created Q&A info pages for parents raising bilingual or trilingual children in Hong Kong and mainland China. We are now replicating findings of Leo’s trilingual development on Leo’s baby brother Louis and exploring the unique role of sibling input in early multilingual development. My PhD student Yuqing Liang is working along this line, and we have a number of outputs in preparation.